Specific fluorescent probe for 8-oxoguanosine

Article abstract of DOI:10.1002/anie.200700671

Marking mutations: Complete discrimination of 8-oxoguanosine (8-oxoG) from other nucleosides has been achieved with a new fluorescent probe, named "8-oxoG-clamp" (see picture). The complex structure between 8-oxoG-clamp and 8-oxoG was confirmed by ^1<

Full text of DOI:10.1002/anie.200700671

Communications
DOI: 10.1002/anie.200700671
Fluorescent Probes
Specific Fluorescent Probe for 8-Oxoguanosine**
Osamu Nakagawa, Sayaka Ono, Zhichun Li, Akira Tsujimoto, and Shigeki Sasaki*
DNA in living organisms suffers from oxidative damage by
reactive oxygen species to form a variety of oxidized nucleo-
sides.[1] 8-Oxoguanosine (8-oxoG) is a representative metab-
olite derived by the oxidation of guanosine (G; Scheme 1),
small differences between G and 8-oxoG. In this paper, the
first example of fluorescent molecules with high specificity
toward 8-oxoG is reported.
We focused on the tricyclic cytosine derivative with
selective affinity toward dG (d: 2’-deoxy) in DNA (G-
clamp, 1).[11] G-clamp (1) is composed of phenoxazine with
an aminoethoxy unit at the end (Scheme 2). It is reported that
Scheme 1. Structures of guanosine and 8-oxoguanosine. dR: b-d-2’-
deoxyribofuranosyl.
and is known to induce G:C to T:A transversion mutations in
DNA (C: cytidine; T: thymidine; A: adenosine).[2] The 8-
oxoG level is regarded as an index of oxidative damage of
cells[3] and is believed to have relevance to some diseases[4]
and aging.[5] Thus, versatile methods for the analysis of 8-
oxoG are of great significance. Until now HPLC-EC,[6]
HPLC/GC-MS,[7] and other techniques[8] have been applied.
Recently, antibodies for 8-oxoG have been developed as a
tool for its detection in an enzyme-linked immunosorbant
assay (ELISA) and the immunohistochemical staining of
tissues or cells.[9]
Small molecules with high specificity to 8-oxoG, especially
those with fluorescence, will find wider utility for the
development of sensors, biological tools, and so on. They
may be more beneficial than antibodies in that they are
potentially applicable to living cells. In this study, we have
attempted to develop fluorescent small molecules for the
recognition of 8-oxoG. In a repair enzyme for 8-oxoG
(hOGG1), the 7-NH group of 8-oxoG is recognized with
one hydrogen bond and no functional group is arranged to
bind to the 8-oxygen atom.[10] The difficulty lies in the
molecular design of small molecules to discriminate such
Scheme 2. Structures of G-clamp and 8-oxoG-clamps. Ac: acetyl; Bz:
benzoyl; Cbz: benzyloxycarbonyl; TBDMS: tert-butyldimethylsilyl.
the specific binding of 1 to dG is due to hydrogen bonding of
the amino group of 1 with the O6 atom of dG, in addition to a
Watson–Crick-type base pairing between the phenoxazine
and the N1, N2, and O6 atoms of dG. In our approach, the
derivatives were designed to have an additional functional
group, benzyloxycarbonyl (in 2), benzoyl (in 3), or acetyl (in
4), at the aminoethoxy terminal and were named “8-oxoG-
clamps” (Scheme 2). It was expected that the carbonyl group
would produce attractive forces with 8-oxoG by hydrogen
bonding with the 7-NH group on one hand or repulsive forces
to the N7 atom of guanosine on the other (Scheme 3).
The functional group of 8-oxoG-clamp derivatives 2–4
was efficiently introduced to the amino group of G-clamp.
The 3’-O and 5’-O positions of the sugar moiety of 2–4 were
protected with TBDMS groups to enhance solubility in
organic solvents.[12]
All probes (1–4) possess intrinsic fluorescence at about
450 nm with excitation at 365 nm; therefore, the binding
properties were investigated by fluorescence quenching.
[*] Dr. O. Nakagawa, S. Ono, Z. Li, A. Tsujimoto, Prof. Dr. S. Sasaki
Graduate School of Pharmaceutical Sciences
Kyushu University
3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582 (Japan)
Fax: (+81)92-642-6615
E-mail: sasaki@phar.kyushu-u.ac.jp
Homepage: http://bioorg.phar.kyushu-u.ac.jp/
[**] This work was supported by a Grant-in-Aid for Scientific Research
(A) from the Japan Society for the Promotion of Science (JSPS). We
are grateful to Dr. A. Takaki, CTO of TAS Project (Fukuoka, Japan),
for financial support and to Y. Tanaka (Faculty of Pharmaceutical
Sciences, Kyushu University, Japan) for technical support with NMR
measurements.
Scheme 3. Design concept and proposed complexation of 8-oxoG-
clamp (2 is shown) with 8-oxoG and dG. dR: 3’O,5’O-di-tert-butyldime-
thylsilyl-b-d-2’-deoxyribofuranosyl.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
4500
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 4500 –4503

Angewandte
Chemie
Titration experiments were performed by adding 3’O,5’O-
disilylated derivatives of dA, dG, dC, dT, or 8-oxo-dG to a
chloroform solution of 8-oxoG-clamps 2–4 buffered with an
organic base and acid. It should be noted that the fluorescence
of 8-oxoG-clamp 2 is effectively quenched by the addition of
8-oxo-dG. In a marked contrast, almost no fluorescence
quenching was observed with dG. Similarly, dA, dC, and dT
did not quench the fluorescence of 2 (Figure 1a--e), which
demonstrates the high selectivity of 2 to 8-oxo-dG. A Job plot
of fluorescence quenching clearly indicated a 1:1 ratio for the
complex between 2 and 8-oxo-dG (Figure 1 f), which was also
showing that the selectivity of 8-oxoG-clamps 2–4 is brought
about by the substitution at the amino group of 1.
When the titration experiments were done with 8-oxoG-
clamp 2 in chloroform under nonbuffered conditions, fluo-
rescence quenching was observed for both 8-oxo-dG with
Ks = 7.1 ” 106 mÀ1 and dG with Ks = 6.9 ” 105 mÀ1 [15]
It has been
.
reported that fluorescence quenching by nucleobases takes
place through a photoinduced electron transfer mechanism
with high dependency on the distance between the chromo-
phores[16] and this may be the case for the fluorescence
quenching observed in this study. Hydrogen bonds between 2
and dG might become weaker in the organic buffer system
than in chloroform because of the polarity of the buffered
solvents, whereas an additional hydrogen bond with the 7-NH
group of 8-oxo-dG might contribute to the retention of
stability of the complex with 2. Thus, the selective fluores-
cence quenching of 2 suggests that, when the two chromo-
phores (2 and 8-oxo-dG or 2 and dG) are forced into
proximity within a tight complex, the fluorescence of 2 is
effectively quenched and that 2 and dG dissociate to some
extent in the organic buffer system to lose the quenching
ability.
In order to characterize the complex structure in detail,
1D and 2D 1H NMR spectra were measured for 2 and 8-oxo-
dG in CD3CN/CDCl3 (v/v, 6:1) or in CD2Cl2 at low temper-
ature. When 2 was added to a solution of 8-oxo-dG, the signals
of complexed and free 8-oxo-dG were observed at different
chemical shifts (Figure 2a), which suggests that complexes are
formed tightly. The signals of the N2H ( H) and N7H ( H)
d
a
protons moved downfield by Dd = 3.8 and 2.2 ppm, respec-
tively. Different regions of the same titration at 08C are
shown in Figure 2b and indicate the downfield shift for the He
proton by complexation. A reverse titration in which 0–2 mm
8-oxo-dG was added to a solution of 2 displayed downfield
shifts for the Hb and Hc protons of 2 by Dd = 2.3 and 2.5 ppm,
respectively (Figure 2c). These downfield shifts indicate that
these NH protons contribute to complex formation by
hydrogen bonds. In the ROESY NMR spectrum of the 1:1
complex of 2 with 8-oxo-dG, clear cross-peaks of Ha–Hb, Hb–
Hc, and Hc–Hd were observed (Figure 2d).[17] In control
experiments under the same conditions, 2 did not show
significant spectrum changes upon mixing with the derivatives
of dA, dT, or dC.[18] The titration with dG indicated lower
binding affinity of 2 with dG than with 8-oxo-dG.[18] These
NMR results strongly support the fact that 2 and 8-oxo-dG
form a complex in the expected manner shown in Figure 2.
Formation of highly ordered multiple hydogen bonds may be
responsible for the high selectivity of 2 for 8-oxo-dG.
Figure 1. a)–e) Fluorescence titration of Cbz-8-oxoG-clamp (2) by 8-
oxo-dG, dG, dA, dC, and dT. Titration conditions: 0–10 mm nucleoside
was added into a solution of 1 mm 2 in CHCl3 with 0.1% dimethylsulf-
oxide (DMSO) buffered with 10 mm triethylamine (TEA) and 2.7 mm
AcOH at 258C. Emission spectra were recorded at lex =365 nm. f) Job
plot with the use of 2 and 8-oxo-dG (total 10 mm). All nucleosides and
2 were used as 3’O,5’O-di-tert-butyldimethylsilyl-2’-deoxynucleosides.
confirmed by ESI-MS measurements.[13] The equilibrium
binding constants between the clamp derivatives 1–4 and 8-
oxo-dG were obtained by these titration experiments and the
results are summarized in Table 1. The binding constant of
Ks = 2.3 ” 106 mÀ1 for 2 is higher than that of a natural G:C
base pair (Ks = 2.0 ” 104 in CDCl3).[14] The fluorescence of the
original G-clamp (1) was quenched almost equally by 8-oxo-
dG (Ks = 7.3 ” 105 mÀ1) and dG (Ks = 7.1 ” 105 mÀ1), clearly
As fluorescent probes are desired to be applicable in
aqueous media, we next examined the fluorescence quench-
ing in water. However, because complexes by hydrogen bonds
are hardly formed in water, no quenching was observed. In a
preliminary investigation for further development, 2 was
solubilized by a detergent (Triton X-100) in water.[19] The
fluorescence intensity of 2 increased dramatically at detergent
concentrations higher than the critical micellar concentration
(cmc), a result indicating that 2 was encapsulated within
micelles. Interestingly, the disilylated nucleoside derivative of
8-oxo-dG exhibited highly selective fluorescence quenching
Table 1: Binding constants of G-clamp (1) and 8-oxoG-clamps 2–4 with
8-oxo-dG and dG in CHCl3 buffered at pH 7.[a]
G-clamp (X)
Ks [mÀ1
]
8-oxo-dG
dG
dA
dC
dT
1 (H)
7.3”105
2.3”106
2.9”105
3.0”105
7.1”105
n.q.
n.q.
n.q.
n.q.
n.q.
n.q.
n.q.
n.q.
n.q.
n.q.
n.q.
n.q.
n.q.
n.q.
2 (Cbz)
3 (Bz)
4 (Ac)
n.q.
[a] Conditions: CHCl3 (0.1% DMSO) buffered with 10 mm TEA and
2.7 mm AcOH, 0–10 mm 8-oxo-dG or dG, 1 mm G-clamp (1) or 8-oxoG-
clamps 2–4 at 258C, excitation 365 nm. n.q.: no fluorescence quenching.
Angew. Chem. Int. Ed. 2007, 46, 4500 –4503
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4501

Communications
ribosyl and ribosyl derivatives, their phosphates, and compo-
nents in nucleic acids. Although further efforts are needed to
develop a useful system for the selective detection of these
metabolites in aqueous media, the results in this study have
suggested great potential for 8-oxoG-clamps as new recog-
nition molecules for 8-oxoguanosine.
In conclusion, we have developed 8-oxoG-clamps as the
first examples of specific fluorescent probes for 8-oxo-dG
with the ability for high discrimination from other nucleo-
1
sides. H NMR analysis has indicated that the complexes are
formed with highly ordered multiple hydrogen bonds in an
expected manner. A preliminary investigation with the use of
a detergent showed that 8-oxoG-clamps might be applicable
in aqueous media. These results indicate high potential for 8-
oxoG-clamps in the development of new analytical systems
for 8-oxoguanosine in aqueous media. The development of
selective fluorescent probes for other oxidized nucleoside
metabolites, such as 8-oxoadenosine, is now in progress.
Received: February 13, 2007
Published online: May 10, 2007
Keywords: DNA damage · fluorescent probes · guanosine ·
.
molecular recognition · nucleosides
[1] a) L. Risom, P. Moller, S. Loft, Mutat. Res. 2005, 592, 119 – 137;
b) A. M. Knaapen, N. Güngör, R. P. F. Schins, P. J. A. Borm, F. J.
Van Schooten, Mutagenesis 2006, 21, 225 – 236; c) A. Valavani-
dis, T. Vlahogianni, M. Dassenakis, M. Scoullos, Ecotoxicol.
Environ. Saf. 2006, 64, 178 – 189; d) B. Halliwell, Biochem. J.
2007, 401, 1 – 11.
Figure 2. a)–c) 1H NMR spectral change during titration and d) a
portion of the ROESY spectrum. a) 0–2 mm 2 was added to 2 mm 8-
oxo-dG at À208C; b) 0–2 mm 2 was added to 2 mm 8-oxo-dG at 08C;
c) 0–2 mm 8-oxo-dG was added to 2 mm 2 at 08C, d) 2 mm mixture of
2 and 8-oxo-dG (1:1) in CD2Cl2 at À108C (mixing time=300 ms).
dR=3’O,5’O-di-tert-butyldimethylsilyl-b-d-2’-deoxyribofuranosyl.
[2] S. Shibutani, M. Takeshita, A. P. Grollaman, Nature 1991, 349,
431 – 434.
[3] a) L. L. Wu, C.-C. Chiou, P.-Y. Chang, J. T. Wu, Clin. Chim. Acta
2004, 339, 1 – 9; b) H. Kasai, Mutat. Res. 1997, 387, 147 – 163;
c) K. P. Cantor, Toxicology 2006, 221, 197 – 204; d) A. R. Collins,
Am. J. Clin. Nutr. 2005, 81, 261S – 267S.
[4] a) P. Jenner, C. W. Olanow, Neurology 1996, 47, S161 – 170; b) S.
Loft, H. E. Poulsen, J. Mol. Med. 1996, 74, 297 – 312; c) I. Dalle-
Donne, R. Rossi, R. Colombo, D. Giustarini, A. Milzani, Clin.
Chem. 2006, 52, 601 – 623.
[5] a) V. A. Bohr, Free Radical Biol. Med. 2002, 32, 804 – 812; b) G.
Barja, Trends Neurosci. 2004, 27, 595 – 600; c) M. F. Beal, Ann.
Neurol. 2005, 58, 495 – 505.
[6] a) J. P. Pouget, T. Douki, M. J. Richard, J. Cadet, Chem. Res.
Toxicol. 2000, 13, 541 – 549; b) A. Siomek, A. Rytarowska, A.
Szaflarska-Poplawska, D. Gackowski, R. Rozalski, T. Dziaman,
M. Czerwionka-Szaflarska, R. Olinski, Carcinogenesis 2005, 27,
405 – 408; c) T. Hofer, A. Y. Seo, M. Prudencio, C. Leeuwen-
burgh, Biol. Chem. 2006, 387, 103 – 113.
of 2, similar to that observed in organic solvents (Figure 3a).
The same quenching experiments were performed in a system
in which dansylamide was cosolubilized by detergent. As
dansylamide emits at 520 nm, a longer wavelength than that
of 2 (440 nm), with the same excitation at 365 nm, it was
expected that quenching of the fluorescence of 2 with the 8-
oxo-dG derivative would induce spectral changes in the
solution.[20] This expectation was successfully proved by the
selective color change with the 8-oxo-dG derivative (Fig-
ure 3b). 8-Oxoguanosine derivatives exist in cells as deoxy-
[7] a) H. Wiseman, H. Kaur, B. Halliwell, Cancer Lett. 1995, 93,
113 – 120; b) M. Dizdaroglu, P. Jaruga, M. Birincioglu, H.
Rodriguez, Free Radical Biol. Med. 2002, 32, 1102 – 1115; c) M.
Lodovici, C. Casalini, R. Cariaggi, L. Michelucci, P. Dolara, Free
Radical Biol. Med. 2000, 28, 13 – 17; d) Y. Nakabeppu, D.
Tsuchimoto, M. Furuichi, K. Sakumi, Free Radical Biol. Med.
2004, 38, 423 – 429; e) Y. Nakae, P. J. Stoward, I. A. Bespalov,
R. J. Melamede, S. S. Wallace, Histochem. Cell Biol. 2005, 124,
335 – 345.
[8] a) M. L. Gielazyn, A. H. Ringwood, W. W. Piegorsch, S. E.
Stancyk, Mutat. Res. 2003, 542, 15 – 22; b) A. Mugweru, B.
Wang, J. Rusling, Anal. Chem. 2004, 76, 5557 – 5563; c) L.
Dennany, R. J. Forster, B. White, M. Smyth, J. F. Rusling, J. Am.
Figure 3. Specific detection of 8-oxo-dG in aqueous media by use of
a) only 2 and b) coexisting 2 and dansylamide. Conditions: a) 20 mm
each for 2, 8-oxo-dG, dG, dA, dC, dT; b) as (a) with 40 mm dansyla-
mide; all solubilized in water with 2.5 mm Triton X-100 at 258C.
Pictures were taken with excitation at 365 nm.
4502 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 4500 –4503

Angewandte
Chemie
Chem. Soc. 2004, 126, 8835 – 8841; d) H. Orimo, N. Mei, S.
Boiteux, Y. Tokura, H. Kasai, J. Radiat. Res. 2004, 45, 455 – 460.
[9] a) R. P. Soultanakis, R. J. Melamede, I. A. Bespalob, S. S.
Wallace, K. B. Beckman, B. N. Ames, D. J. Taatjes, W. M. W.
Janssen-Heininger, Free Radical Biol. Med. 2000, 28, 987 – 998;
b) C. C. Chiou, P. Y. Chang, E. C. Chan, T. L. Wu, K. C. Tsao,
J. T. Wu, Clin. Chim. Acta 2003, 334, 87 – 94; c) N. Machella, F.
Regoli, A. Cambria, R. M. Santella, Mar. Environ. Res. 2004, 58,
725 – 729; d) Y. Nakae, P. J. Stoward, I. A. Bespalov, R. J.
Melamede, S. S. Wallace, Histochem. Cell Biol. 2005, 124, 335 –
345.
istry 2002, 41, 1323 – 1327; d) C. J. Wilds, M. A. Maier, V.
Tereshko, M. Manoharan, M. Egli, Angew. Chem. 2002, 114,
123 – 125; Angew. Chem. Int. Ed. 2002, 41, 115 – 117; e) S. C.
Holmes, A. A. Arzumanov, M. J. Gait, Nucleic Acids Res. 2003,
31, 2759 – 2768.
[12] See Scheme S1 in the Supporting Information.
[13] See Figure S1 in the Supporting Information.
[14] P. M. Gannett, T. P. Sura, Chem. Res. Toxicol. 1993, 6, 690 – 700.
[15] See Table S1 in the Supporting Information.
[16] M. Torimura, S. Kurata, K. Yamada, T. Yokomaku, Y. Kamagata,
T. Kanagawa, R. Kurane, Anal. Sci. 2001, 17, 155 – 160.
[17] See Figure S2 in the Supporting Information.
[18] See Figure S3 in the Supporting Information.
[19] Y. Diaz Fernandez, A. P. Gramatges, V. Amendola, F. Foti, C.
Mangano, P. Pallavicini, S. Patroni, Chem. Commun. 2004, 1650 –
1651.
[10] S. D. Bruner, D. P. G. Norman, G. L. Verdine, Nature 2000, 403,
859 – 866.
[11] a) K.-Y. Lin, M. D. Matteucci, J. Am. Chem. Soc. 1998, 120,
8531 – 8532; b) W. M. Flanagan, J. J. Wolf, P. Olson, D. Grant, K.-
Y. Lin, R. W. Wagner, M. D. Matteucci, Proc. Natl. Acad. Sci.
USA 1999, 96, 3513 – 3518; c) M. A. Maier, J. M. Leeds, G.
Balow, R. H. Springer, R. Bharadwaj, M. Manoharan, Biochem-
[20] See Figures S4 and S5 in the Supporting Information.
Angew. Chem. Int. Ed. 2007, 46, 4500 –4503
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4503